Desired Accuracy Research Articles

A light-weight skeleton human action recognition model with knowledge distillation for edge intelligent surveillance applications

Skeleton based human action recognition has evolved as one of the most important applications in multimedia IoT system. However, it requires extensive computation resource including high performance computing unites and large memory to train a deep mode with large number of parameters, which seriously limits it effectiveness and efficiency for edge intelligence multimedia IoT applications. In this paper, a knowledge distillation based light-weight deep model is proposed for skeleton human action recognition to meet the edge multimedia IoT applications. It can get competitive recognition performance in terms of learning accuracy for combination of AI model and edge surveillance equipment. On the one hand, to achieve desirable accuracy, we propose a deep pose-transition image representation method based on two-stream spatial–temporal architecture, which can mine the hidden features of color texture images in spatial and temporal domain, and fuse them for comprehensive discrimination before final classification. On the other hand, to increase the transfer learning ability to the student model on the edge device, we use tucker decomposition to weak the teacher model during knowledge transfer learning process. Finally, in order to validate the effectiveness of our proposal, we conducted extensive experiments to evaluate the proposed approach. The experimental results demonstrate that our proposal can realize deep model miniaturization to meet the requirement of edge multimedia IoT system and achieve the competitive performance.

Prescribed-Time Formation Control for a Class of Multiagent Systems via Fuzzy Reinforcement Learning

This paper concerns optimal prescribed-time forma- tion control for a class of nonlinear multi-agent systems (MASs). Optimal control depends on the solution of the Hamilton-Jacobi- Bellman equation, which is hard to be calculated directly due to its inherent nonlinearity. To overcome this difficulty, the rein- forcement learning strategy with fuzzy logic systems is proposed, in which identifier, actor, and critic are used to estimate unknown nonlinear dynamics, implement control behavior, and evaluate system performance, respectively. Different from the existing optimal control algorithms, a new performance index function considering formation error cost and control input energy cost is constructed to achieve optimal formation control of MASs within a prescribed time. The presented control strategy can ensure that the formation error converges to the desired accuracy within a prescribed time. Finally, the validity of the presented strategy is verified via a simulation example.

Load-current demand forecasting in substations of urban railway lines

Load-current forecasting is important in the operation and energy management of urban railway lines. It helps control strategies to manage and distribute energy optimally, thereby saving energy and reducing voltage fluctuations. This paper presents a method to predict the traction current at the busbar of a substation using the supervised machine learning algorithm. Because the traction power load is supplied from both adjacent traction power stations, and the energy exchange process between trains also takes place during work, the input data are selected to combine the value history of busbar currents and feeder currents. Besides, the neural network configuration and the number of training cycles in the estimated model can be adjusted to achieve the desired accuracy. The proposed method was tested and adjusted based on the actual operation data at the Lang traction station on 24 June, 2022. The estimated results compared with measurement data from the supervisory control and data acquisition (SCADA) have proven that the largest absolute error is no more than 5 (A). The maximum relative error is not more than 0.005, and the mean squared error does not exceed 0.01 over the whole operating time of a day from 4h45 to 22h45.

Inexact tensor methods and their application to stochastic convex optimization

We propose general non-accelerated [The results for non-accelerated methods first appeared in December 2020 in the preprint (A. Agafonov, D. Kamzolov, P. Dvurechensky, and A. Gasnikov, Inexact tensor methods and their application to stochastic convex optimization, preprint 2020. arXiv:2012.15636)] and accelerated tensor methods under inexact information on the derivatives of the objective, analyse their convergence rate. Further, we provide conditions for the inexactness in each derivative that is sufficient for each algorithm to achieve a desired accuracy. As a corollary, we propose stochastic tensor methods for convex optimization and obtain sufficient mini-batch sizes for each derivative.

Adaptive VEM for variable data: convergence and optimality

Abstract We design an adaptive virtual element method (AVEM) of lowest order over triangular meshes with hanging nodes in 2d, which are treated as polygons. AVEM hinges on the stabilization-free a posteriori error estimators recently derived in Beirão da Veiga et al. (2023, Adaptive VEM: stabilization-free a posteriori error analysis and contraction property. SIAM J. Numer. Anal., 61, 457–494). The crucial property, which also plays a central role in this paper, is that the stabilization term can be made arbitrarily small relative to the a posteriori error estimators upon increasing the stabilization parameter. Our AVEM concatenates two modules, GALERKIN and DATA. The former deals with piecewise constant data and is shown in the above article to be a contraction between consecutive iterates. The latter approximates general data by piecewise constants to a desired accuracy. AVEM is shown to be convergent and quasi-optimal, in terms of error decay versus degrees of freedom, for solutions and data belonging to appropriate approximation classes. Numerical experiments illustrate the interplay between these two modules and provide computational evidence of optimality.

Textile-Sensing Wearable Systems for Continuous Motion Angle Estimation: A Systematic Review

Textile sensors have demonstrated significant potential in next-generation wearable systems due to their excellent performance and unobtrusive nature. By building specialized sensing networks and algorithms, textile-based wearable systems can estimate the continuous motion angles of human joints with desirable accuracies. This article offers a systematic review aimed at identifying key challenges in this field and encouraging further applications of textile strain sensor networks within the human–computer interaction (HCI) community. To achieve this, we conducted an exhaustive literature search across four major databases: IEEE Xplore, PubMed, Scopus, and Web of Science, spanning from January 2016 to August 2023. Applying inclusion and exclusion criteria, we narrowed down 2684 results to a total of 24 relevant papers. To analyze these studies, we proposed a framework that incorporates both technical aspects – such as textile strain sensors, sensor placement, algorithms, and technical evaluations – and contextual factors like target users, wearability, and application scenarios. Our analysis uncovered two critical research gaps: First, it exists an incongruity between the development of textile-based wearables and the advancements in textile sensors. Second, there is a noticeable absence of contextual design considerations in this specific domain. To address these issues, we offer discussions and recommendations from three perspectives: 1) enhancing the robustness of textile-sensing networks, 2) improving wearability, and 3) expanding application scenarios.

A multi-fidelity surrogate model based on design variable correlations

Multi-fidelity surrogate (MFS) models have garnered significant attention in the field of engineering optimization due to their ability to attain the desired accuracy at a reduced cost. However, most previous MFS models employed a single scaling factor for the global design space, which posed challenges in adaptively adjusting the scaling factor within local design spaces. To address this issue, this paper proposes a novel MFS model based on design variable correlations (MFS-DVC). MFS-DVC introduces local characteristics to the scaling factors by leveraging correlations between design variables, enabling adaptive adjustments of the scaling factors at different positions within the design space. Moreover, MFS-DVC offers a more comprehensive exploration of relationships and information among high-fidelity (HF) data points by utilizing the correlations between design variables. This enhancement contributes to improving the accuracy and robustness of the model. The performance of MFS-DVC was evaluated by comparing it with three benchmark MFS models and two single-fidelity surrogate models on 22 test functions and one engineering problem involving a hydraulic press. Additionally, the cost ratio and combination of HF and low-fidelity samples were studied to assess their impact on the performance of the MFS-DVC model. The results demonstrate that MFS-DVC consistently delivers competitive performance in terms of prediction accuracy and robustness.

Stable approximation of Helmholtz solutions in the disk by evanescent plane waves

Superpositions of plane waves are known to approximate well the solutions of the Helmholtz equation. Their use in discretizations is typical of Trefftz methods for Helmholtz problems, aiming to achieve high accuracy with a small number of degrees of freedom. However, Trefftz methods lead to ill-conditioned linear systems, and it is often impossible to obtain the desired accuracy in floating-point arithmetic. In this paper we show that a judicious choice of plane waves can ensure high-accuracy solutions in a numerically stable way, in spite of having to solve such ill-conditioned systems. Numerical accuracy of plane wave methods is linked not only to the approximation space, but also to the size of the coefficients in the plane wave expansion. We show that the use of plane waves can lead to exponentially large coefficients, regardless of the orientations and the number of plane waves, and this causes numerical instability. We prove that all Helmholtz fields are continuous superposition of evanescent plane waves, i.e., plane waves with complex propagation vectors associated with exponential decay, and show that this leads to bounded representations. We provide a constructive scheme to select a set of real and complex-valued propagation vectors numerically. This results in an explicit selection of plane waves and an associated Trefftz method that achieves accuracy and stability. The theoretical analysis is provided for a two-dimensional domain with circular shape. However, the principles are general and we conclude the paper with a numerical experiment demonstrating practical applicability also for polygonal domains.

Digital holographic interferometry method for tracking detector modules displacement

In high energy particle physics scattering experiments, the precision of the reconstructed particle tracks can be fundamental. For this reason, a method for detecting the displacement of tracking detector modules is developed. The modules are silicon planes mounted on a frame and used in the MUonE project, which aims at a precision measurement of the scattering angle of elastic muon-electron scattering. From the scattering angle, the hadronic contribution to the anomalous magnetic moment of the muon is extracted. To achieve the desired accuracy, the position of the tracking detector planes must be monitored. The allowable relative displacements must be less than 10 μ m. To meet the specifications and to monitor as large an area of the detector as possible, a digital holographic interferometer was developed. It is based on a novel lens-less design in off-axis holographic geometry. Light from a fiber-coupled laser source is split by a fiber beam splitter, with one output used to illuminate the detector plane and the other for the reference beam. The two beams produce an interference pattern on a CMOS image sensor. To obtain relative displacement information, successive images are superimposed on an initial reference image and reconstructed by solving the Rayleigh-Sommerfeld diffraction integral taking into account the spherical wavefronts of the beams. The interference fringes that appear in the reconstructed holographic image provide a measure of the relative displacement of the detector plane compared to the initial position. The performance of the reconstruction method used was verified with the proposed setup at a real tracking station.

A Tutorial on Retroreflectors and Arrays Used in Satellite and Lunar Laser Ranging

The present paper discusses the basics of retroreflector theory and the manner in which they are combined in arrays to service the laser tracking of artificial satellites and the Moon. We begin with a discussion of the relative advantages and disadvantages of solid versus hollow cube corners and the functional dependence of their optical cross-sections and far-field patterns on cube diameter. Because of velocity aberration effects, the design of an array for a particular space mission depends on many factors, including the desired range accuracy and the satellite’s orbital altitude, velocity, and pass geometry relative to the tracking station. This generally requires the individual retroreflectors in the array to be “spoiled” by perturbing one or more of the 90-degree angles that define a perfect cube corner, or alternatively, by adding a curved surface to a hollow cube. In order to obtain adequate return signal strengths from all points along the satellite path, the rotational orientation of the retroreflectors within the array may need to be varied or “clocked”. Possible approaches to developing millimeter-accuracy arrays with both large cross-sections and ultrashort satellite signatures are discussed, as are new designs proposed to replace aging reflectors on the Moon. Finally, we briefly discuss methods for laser ranging beyond the Moon.

Surrogate Lagrangian Relaxation: A Path to Retrain-Free Deep Neural Network Pruning

Network pruning is a widely used technique to reduce computation cost and model size for deep neural networks. However, the typical three-stage pipeline (i.e., training, pruning, and retraining (fine-tuning)) significantly increases the overall training time. In this article, we develop a systematic weight-pruning optimization approach based on surrogate Lagrangian relaxation (SLR), which is tailored to overcome difficulties caused by the discrete nature of the weight-pruning problem. We further prove that our method ensures fast convergence of the model compression problem, and the convergence of the SLR is accelerated by using quadratic penalties. Model parameters obtained by SLR during the training phase are much closer to their optimal values as compared to those obtained by other state-of-the-art methods. We evaluate our method on image classification tasks using CIFAR-10 and ImageNet with state-of-the-art multi-layer perceptron based networks such as MLP-Mixer; attention-based networks such as Swin Transformer; and convolutional neural network based models such as VGG-16, ResNet-18, ResNet-50, ResNet-110, and MobileNetV2. We also evaluate object detection and segmentation tasks on COCO, the KITTI benchmark, and the TuSimple lane detection dataset using a variety of models. Experimental results demonstrate that our SLR-based weight-pruning optimization approach achieves a higher compression rate than state-of-the-art methods under the same accuracy requirement and also can achieve higher accuracy under the same compression rate requirement. Under classification tasks, our SLR approach converges to the desired accuracy × faster on both of the datasets. Under object detection and segmentation tasks, SLR also converges 2× faster to the desired accuracy. Further, our SLR achieves high model accuracy even at the hardpruning stage without retraining, which reduces the traditional three-stage pruning into a two-stage process. Given a limited budget of retraining epochs, our approach quickly recovers the model’s accuracy.

Taxicab Metric of a Different Kind

Summary Imagine being tasked with measuring the distance between two points in a city, but you can only use a taxicab’s odometer which displays only whole-number values (the integer part) of its distance traveled so far—such is the display of modern digital measuring devices. This coarse measurement between two points might be called, tongue in cheek, a taxicab metric, but clearly of a different sort from the taxicab metric on a discrete grid that we encounter in combinatorics. We will explore how we might apply probability to our taxicab’s inherent randomness, and along with limit theorems, indeed measure arbitrary distances with any desired accuracy.

Impact of Arrhenius energy and irregular heat absorption on generalized second grade fluid MHD flow over nonlinear elongating surface with thermal radiation and Cattaneo–Christov heat flux theory

In this paper, the free convection flow phenomenon on magnetized second-grade nanofluid over a nonlinear elongating surface has been modeled under the assumptions of generalized Fick’s and Cattaneo–Christov heat flux relations. The non-uniform heat source and sink and thermal radiation effects are used in the energy equation with the Buongiorno nanofluid model. The effect on activation energy and chemical processes is also examined, making the current numerical scrutinization innovative. The current mathematical model begins with partial derivative equations (PDEs), which are then transformed into ordinary derivative equations (ODEs) using similarity transformations. The outcomes are obtained using a combination of the finite-difference scheme with the collocation method, approximated up to desirable accuracy, and the effects of various parameters on the modified second-grade nanofluid are discussed using tables and graphs. The temperature distribution for the thermal boundary layer is raised with nonlinear thermal radiation and Brownian motion parameters. The present thorough analysis identified several technical and industrial applications, including Biomedical applications in cancer treatment, electrical wire manufacturing, oil friction control in pipelines, manufacturing processes, and aerodynamic expulsion utilizations.

Matrix effect in bioanalytical assay development using supercritical fluid chromatography-mass spectrometry.

Matrix effect (ME) is commonly caused by coelution of compounds with target analytes, resulting in either suppression or enhancement of analyte ionization. Thus, to achieve the desired accuracy, precision, and sensitivity, ME needs to be evaluated and controlled during bioanalytical method development. As the application of supercritical fluid chromatography-mass spectrometry (SFC-MS) for analysis of biological samples has increased, ME using SFC-MS has also been investigated with a focus on the difference in ME in SFC-MS compared to other chromatographic techniques used for achiral separation in biological samples. Here, we provide a summary of the status of ME evaluation and mitigation in SFC-MS methods. This review presents an overview of the phenomenon of ME and methods for evaluating ME in bioanalysis. Next, the factors that can impact ME in SFC-MS-based bioanalytical methods are discussed in detail with an emphasis on SFC. A literature review of the evaluation of ME in targeted bioanalytical methods using SFC-MS is included at the end. Robust instrumentation, effective sample preparation, and superb separation selectivity are the foundations of reliable analytical methods as well as the ability to mitigate detrimental ME in SFC-MS methods.

Control of magnetic dissipation and radiation on an unsteady stagnation point nanofluid flow: A numerical approach

This study investigates two-dimensional time-dependent stagnation point nanofluid flow over a permeable stretching or sinking sheet. The electrically conducting fluid allows for the interaction of a transverse magnetic field along with magnetic dissipation and thermal radiation to enrich the flow phenomena. The use of nanofluids is becoming increasingly important in a variety of industrial applications, as well as in engineering and biomedicine. The process of peristaltic pumping, the flow of blood in the artery, the drug delivery process, and besides that, the cooling of electronic devices for the better shape of the product at the final stage of production, the use of nanofluid, are all important. The transformed dimensionless governing equations are obtained by the implementation of various transformation rules, and the model is designed by using different thermophysical properties such as viscosity, conductivity, etc. Finally, the numerical technique RK-fourth-order method is deployed to solve the set of equations and the iteration process is continued up to the desired accuracy of 10[Formula: see text]. The validation of the current result is shown with the earlier investigation in particular cases, and further, the behavior of the physical parameters is presented through graphs. It is observed that the particle concentration has a greater impact in enhancing the velocity distribution for the variation of suction/injection. The behavior of the magnetic parameter vis-à-vis the unsteadiness parameter and the thermal buoyancy presents their greater contribution in enhancing the magnitude of nanofluid velocity.

Determination of tobacco-specific nitrosamines in tobacco and mainstream cigarette smoke using one-step clean-up coupled with liquid chromatography- tandem mass spectrometry

A method for determining tobacco-specific nitrosamines (TSNAs) in tobacco and cigarette smoke using liquid chromatography-tandem mass spectrometry was established. The established method amended the deficiencies that exist in current mainstream methods. In this method, TSNAs in tobacco and cigarette smoke were extracted by water. The aqueous extract was then extracted by dichloromethane, and the extract could be analyzed by liquid chromatography-tandem mass spectrometry after a solvent replacement. This method was used to analyze flue-cured tobacco samples, and the response of the target compounds was about 10 times higher than that of the ammonium acetate extraction method. When analyzing cigarette smoke samples, the response strength and chromatographic peak purity of the target compounds were also significantly improved. The proposed method exhibited good linearities for both tobacco and cigarette smoke samples (r2 > 0.99). The limits of detection (LODs) for tobacco and cigarette smoke samples were 0.2–1.0 ng/g and 0.1–0.3 ng/cigarette, respectively. Additionally, this method exhibited desirable accuracy and precision. The TSNAs recovery values from tobacco and cigarette smoke samples ranged from 95.7 % to 107.7 % with inter- and intra-day relative standard deviations (RSDs) of less than 7.4 %. This method is simple, effective, and has wide adaptability. It is a useful upgrade to the existing methods for analyzing TSNAs in tobacco and cigarette smoke.

Energy and Distribution-Aware Cooperative Clustering Algorithm in Internet of Things (IoT)-Based Federated Learning

In Internet of Things (IoT)-based federated learning (FL), if IoT devices are located far from the base station (BS), they consume lots of energy to transmit the updated parameters to BS, which can cause energy depletion of IoT devices. To mitigate this problem, IoT devices can be clustered and the clustered IoT devices transmit the updated parameters to their cluster headers (CHs) instead of BS. However, if the aggregated data distribution in each cluster is non-independent and identically distributed (non-IID), the desired accuracy of the model cannot be achieved. In this paper, we propose an energy and distribution-aware cooperative clustering algorithm (EDA-CCA) where several IoT devices with a sufficient energy level are selected as CHs. By considering the distance to these CHs and BS and the data distribution of IoT devices, other IoT devices are clustered, and then a hierarchical parameter aggregation (i.e., sequential aggregation within each cluster and synchronous aggregation between CHs and FL server) is conducted. Evaluation results demonstrate that EDA-CCA can make the model having the desired accuracy with the lowest energy consumption among the comparison schemes.

Research on land cover classification of multi-source remote sensing data based on improved U-net network

In recent years, remote sensing images of various types have found widespread applications in resource exploration, environmental protection, and land cover classification. However, relying solely on a single optical or synthetic aperture radar (SAR) image as the data source for land cover classification studies may not suffice to achieve the desired accuracy in ground information monitoring. One widely employed neural network for remote sensing image land cover classification is the U-Net network, which is a classical semantic segmentation network. Nonetheless, the U-Net network has limitations such as poor classification accuracy, misclassification and omission of small-area terrains, and a large number of network parameters. To address these challenges, this research paper proposes an improved approach that combines both optical and SAR images in bands for land cover classification and enhances the U-Net network. The approach incorporates several modifications to the network architecture. Firstly, the encoder-decoder framework serves as the backbone terrain-extraction network. Additionally, a convolutional block attention mechanism is introduced in the terrain extraction stage. Instead of pooling layers, convolutions with a step size of 2 are utilized, and the Leaky ReLU function is employed as the network's activation function. This design offers several advantages: it enhances the network's ability to capture terrain characteristics from both spatial and channel dimensions, resolves the loss of terrain map information while reducing network parameters, and ensures non-zero gradients during the training process. The effectiveness of the proposed method is evaluated through land cover classification experiments conducted on optical, SAR, and combined optical and SAR datasets. The results demonstrate that our method achieves classification accuracies of 0.8905, 0.8609, and 0.908 on the three datasets, respectively, with corresponding mIoU values of 0.8104, 0.7804, and 0.8667. Compared to the traditional U-Net network, our method exhibits improvements in both classification accuracy and mIoU to a certain extent.

PMVT: a lightweight vision transformer for plant disease identification on mobile devices.

Due to the constraints of agricultural computing resources and the diversity of plant diseases, it is challenging to achieve the desired accuracy rate while keeping the network lightweight. In this paper, we proposed a computationally efficient deep learning architecture based on the mobile vision transformer (MobileViT) for real-time detection of plant diseases, which we called plant-based MobileViT (PMVT). Our proposed model was designed to be highly accurate and low-cost, making it suitable for deployment on mobile devices with limited resources. Specifically, we replaced the convolution block in MobileViT with an inverted residual structure that employs a 7×7 convolution kernel to effectively model long-distance dependencies between different leaves in plant disease images. Furthermore, inspired by the concept of multi-level attention in computer vision tasks, we integrated a convolutional block attention module (CBAM) into the standard ViT encoder. This integration allows the network to effectively avoid irrelevant information and focus on essential features. The PMVT network achieves reduced parameter counts compared to alternative networks on various mobile devices while maintaining high accuracy across different vision tasks. Extensive experiments on multiple agricultural datasets, including wheat, coffee, and rice, demonstrate that the proposed method outperforms the current best lightweight and heavyweight models. On the wheat dataset, PMVT achieves the highest accuracy of 93.6% using approximately 0.98 million (M) parameters. This accuracy is 1.6% higher than that of MobileNetV3. Under the same parameters, PMVT achieved an accuracy of 85.4% on the coffee dataset, surpassing SqueezeNet by 2.3%. Furthermore, out method achieved an accuracy of 93.1% on the rice dataset, surpassing MobileNetV3 by 3.4%. Additionally, we developed a plant disease diagnosis app and successfully used the trained PMVT model to identify plant disease in different scenarios.

Development and validation of a competitive risk model in patients with rectal cancer: based on SEER database

BackgroundRectal cancer is one of the most common malignancies. To predict the specific mortality risk of rectal cancer patients, we constructed a predictive nomogram based on a competing risk model.MethodsThe information on rectal cancer patients was extracted from the SEER database. Traditional survival analysis and specific death analysis were performed separately on the data.ResultsThe present study included 23,680 patients, with 16,580 in the training set and 7100 in the validation set. The specific mortality rate calculated by the competing risk model was lower than that of the traditional survival analysis. Age, Marriage, Race, Sex, ICD-O-3Hist/Behav, Grade, AJCC stage, T stage, N stage, Surgery, Examined LN, RX SUMM-SURG OTH, Chemotherapy, CEA, Deposits, Regional nodes positive, Brain, Bone, Liver, Lung, Tumor size, and Malignant were independent influencing factors of specific death. The overall C statistic of the model in the training set was 0.821 (Se = 0.001), and the areas under the ROC curve for cancer-specific survival (CSS) at 1, 3, and 5 years were 0.842, 0.830, and 0.812, respectively. The overall C statistic of the model in the validation set was 0.829 (Se = 0.002), and the areas under the ROC curve for CSS at 1, 3, and 5 years were 0.851, 0.836, and 0.813, respectively.ConclusionsThe predictive nomogram based on a competing risk model for time-specific mortality in patients with rectal cancer has very desirable accuracy. Thus, the application of the predictive nomogram in clinical practice can help physicians make clinical decisions and follow-up strategies.